760 likes | 986 Views
William Stallings Data and Computer Communications. Chapter 7 Data Link Control. Flow Control. Ensuring the sending entity does not overwhelm the receiving entity Preventing buffer overflow Transmission time Time taken to emit all bits into medium Propagation time
E N D
William StallingsData and Computer Communications Chapter 7 Data Link Control
Flow Control • Ensuring the sending entity does not overwhelm the receiving entity • Preventing buffer overflow • Transmission time • Time taken to emit all bits into medium • Propagation time • Time for a bit to traverse the link
Stop and Wait • Source transmits frame • Destination receives frame and replies with acknowledgement • Source waits for ACK before sending next frame • Destination can stop flow by not send ACK • Works well for a few large frames
Fragmentation • Large block of data may be split into small frames • Limited buffer size • Errors detected sooner (when whole frame received) • On error, retransmission of smaller frames is needed • Prevents one station occupying medium for long periods • Stop and wait becomes inadequate
Sliding Windows Flow Control • Allow multiple frames to be in transit • Receiver has buffer W long • Transmitter can send up to W frames without ACK • Each frame is numbered • ACK includes number of next frame expected • Sequence number bounded by size of field (k) • Frames are numbered modulo 2k
Sliding Window Enhancements • Receiver can acknowledge frames without permitting further transmission (Receive Not Ready) • Must send a normal acknowledge to resume • If duplex, use piggybacking • Acknowledgment is sent with data in same frame • If no data to send, use acknowledgement frame • If data but no acknowledgement to send, send last acknowledgement number again, or have ACK valid flag (TCP)
Error Detection • Transmission system results in errors • Pb: Probability of a single bit error (BER) • P1: Probability that a frame arrives with no bit errors • Pf: Probability that a frame arrives with one or more bit errors; Pf=1- P1 • Pfd: Probability that a faulty frame is detected • P2: Probability that a frame arrives with one or more undetected bit errors; P2= Pf x (1- Pfd) • P3: Probability that a frame arrives with one or more detected bit errors but no undetected bit errors; P3= Pf x Pfd • With no error detection P3=0
Error Detection • Assuming Pb is constant and independent of bit position and error detection is not employed • P1=(1- Pb)F; F is number of bits in a frame • P2=1- P1 • Objective for ISDN connections is that BER on a 64 Kbps channel < 10-6 on at least 90% of observed 1-minute interval • Assume frame length=1000 bits, 64 Kbps channel, one frame with undetected bit error per day • Number of transmitted frames in a day=5.529x106 • Desired frame error rate P2=1/(5.529x106)=0.18x10-6 • Assume Pb= 10-6, P1=(1- 0.000001)1000=0.999; P2= 10-3 • Error detection techniques are necessary
Error Detection • Additional bits added by transmitter for error detection code
Error Detection • Parity • Value of parity bit is such that character has even (even parity) or odd (odd parity) number of ones • Even number of bit errors goes undetected • Cyclic Redundancy Check • For a block of k bits, transmitter generates n bit sequence called frame check sequence (FCS) • Transmit k+n bits which is exactly divisible by some number • Receive divides frame by that number • If no remainder, assume no error
Cyclic Redundancy Check • M= K-bit message • F=n-bit FCS • T=(K+n)-bit frame to be transmitted, n<K • P=pattern of n+1 bits; predetermined divisor • Requirement is that T/P has no remainder
Cyclic Redundancy Check - Example • Message M = 1010001101 (10 bits) • Pattern P = 110101 (6 bits) • FCS R = 01110 (5 bits)
Cyclic Redundancy Check - Polynomials • Values expressed as polynomials in a dummy variable X with binary coefficients • For M=110011, M(X)=X5+X4+X+1 • For P=11001, P(X)=X4+X3+1 • CRC process:
Cyclic Redundancy Check – Error Detection • Errors in an (n+k)-bit frame represented by an (n+k)-bit field with 1’s in each error position • T=Transmitted frame • E=Error pattern with 1’s in error positions • Tr=Received frame • Error is not detected by receiver iff Tr is divisible by P, => E is divisible by P
Cyclic Redundancy Check – Error Detection • The following errors are not divisible by a suitably chosen p(x) & are detectable: • All single bit errors • All double bit errors, as long as P(X) has at least three 1’s • Any odd number of errors as long as P(X) contains a factor of (1+X) • Any burst error with a length < length of P(X) • Most larger burst errors
Cyclic Redundancy Check – Error Detection • Assuming r=length of FCS • For a burst error of length r+1, probability of an undetected error=1/2r-1 • For longerburst,probability of an undetected error=1/2r • Four versions of P(X) are widely used • CRC-12: =X12+X11+X3+X2+X+1 • CRC-16: =X16+X15+X2+1 • CRC-CCITT: =X16+X12+X5+1 • CRC-32: =X32+X26+X23+X22+X16+X12+X11+X10+X8 +X7+X5+X4+X2+X+1
Error Control • Detection and correction of errors • Types of errors • Lost frames • Damaged frames • Automatic repeat request (ARQ) • Error detection • Positive acknowledgment • Retransmission after timeout • Negative acknowledgement and retransmission
Automatic Repeat Request (ARQ) • ARQ turns unreliable data link into a reliable one • Standardized ARQ versions • Stop and wait • Go back N • Selective reject (selective retransmission)
Stop and Wait • Source transmits single frame • Wait for ACK • If received frame damaged, discard it • Transmitter has timeout • If no ACK within timeout, retransmit • If ACK damaged,transmitter will not recognize it • Transmitter will retransmit • Receive gets two copies of frame • Use ACK0 and ACK1
Stop and Wait • Simple • Inefficient
Go Back N • Based on sliding window • If no error, ACK as usual with next frame expected • Use window to control number of outstanding frames • If error, reply with rejection • Discard that frame and all future frames until error frame received correctly • Transmitter must go back and retransmit that frame and all subsequent frames
Go Back N - Damaged Frame • Receiver detects error in frame i • Receiver sends Reject i • Transmitter gets Reject i • Transmitter retransmits frame i and all subsequent frames
Go Back N - Lost Frame: Case1 • Frame i lost • Transmitter sends i+1 • Receiver gets frame i+1 out of sequence • Receiver send Reject i • Transmitter goes back to frame i and retransmits
Go Back N - Lost Frame: Case2 • Frame i lost and no additional frame sent • Receiver gets nothing and returns neither acknowledgement nor rejection • Transmitter times out and sends acknowledgement frame with P bit set to 1 • Receiver interprets this as command which it acknowledges with the number of the next frame it expects (frame i ) • Transmitter then retransmits frame i
Go Back N - Damaged Acknowledgement • Receiver gets frame i and send ACK (i+1) which is lost • Acknowledgements are cumulative, so next ACK (i+n) may arrive before transmitter times out on frame i • If transmitter times out, it sends acknowledgement with P bit set as before • This can be repeated a number of times before a reset procedure is initiated
Go Back N - Damaged Rejection • Receiver sends Reject i that is lost • Transmitter times out and sends acknowledgement frame with P bit set to 1 • Receiver interprets this as command which it acknowledges with the number of the next frame it expects (frame i ) • Transmitter then retransmits frame i
Selective Reject • Also called selective retransmission • Only rejected frames are retransmitted • Subsequent frames are accepted by the receiver and buffered • Minimizes retransmission • More complex login in transmitter & receiver
Maximum Window Size • For a k-bit sequence with 2k sequence numbers, maximum window size is 2k-1. • Data is being exchanged in both directions • Station B sends piggybacked acknowledgement with data frames sent to A • Assume a 3-bit sequence number • Station A sends frame 0 and gets back an RR 1 • Station A then sends frames 1,2,3,4,5,6,7,0 and gets another RR 1 • This could mean either • All eight frames received correctly and acknowledged by RR1 • All eight frames were damaged or lost and receiver is repeating its previous RR 1.
Stop-and-Wait Performance • Let T be the time to send a frame and receive an acknowledgment • T=tframe+tprop+tproc+tack+tprop+ tproc • Assume tprocis negligibleand tack is very small compared to tframe • T=tframe+ 2tprop • Utilization or efficiency of line
Stop-and-Wait Performance • Let d be the distance link, V the velocity of propagation, L the frame length, and R the data rate • V is the speed of light • For unguided media V=3x108 m/s • For optical fiber and copper media V= 2x108 m/s
Stop-and-Wait Performance Examples • Consider a WAN using ATM with two stations 1000 Km apart • In ATM, frame size is 53 bytes (424 bits) • Assume data rate of 155.52 Mbps • Transmission time=424/(155.52x106)=2.7x10-6 • Assuming optical fiber link, propagation time=(106 meters)/(2x108 m/s)=0.5x10-2 • a=(0.5x10-2)/(2.7x10-6)=1851 • Efficiency=1/3703=0.00027
Stop-and-Wait Performance Examples • Assume a LAN with a frame size of 1000 bits, and data rate of 10 Mbps • For distance of 0.1 km • propagation time=(102 meters)/(2x108 m/s)=0.5x10-6 • Transmission time=1000/(107)=10-4 • a=(0.5x10-6)/(10-4)=0.005 • Efficiency=1/1.01=0.99 • For distance of 10 km • propagation time=(104 meters)/(2x108 m/s)=0.5x10-4 • Transmission time=1000/(107)=10-4 • a=(0.5x10-4)/(10-4)=0. 5 • Efficiency=1/2=0.5
Stop-and-Wait Performance Examples • Consider digital data transmission via modem over a voice-grade line with data rate equal to 56 kbps. • Consider a 1000-bit frame. • The link distance can range from a few tens of meters to thousands of kilometers. • For a short distance d = 1000 m, then a = (1000 m / 2 x 108 m/s) /(1000 bit / 56,000 bps) = 2.8 x 10-4, and utilization is effectively 1.0. • For a long distance d=5000 km, (5x106 m / 2 x 108 m/s) /(1000 bit / 56,000 bps) = 1.4, and utilization is 0.26.
Error-Free Sliding Window Flow Control Performance • Case 1: w >= 2a+1 • Acknowledgment for frame 1 reaches A before A has exhausted its window • A can transmit continuously with no pause, and efficiency =1. • Case 1: w < 2a+1 • A exhausts its window at t=W and cannot send additional frames until t=2a+1
Stop-and-Wait ARQ • For error-free link • In the presence of errors • Nr is the expected number of transmissions of a frame • P is the probability that a single frame is in error
Selective Reject ARQ Performance • For error-free link • In presence of errors
Go-Back-N ARQ Performance • Each error generates a requirement to retransmit K frames rather than just one frame • f(i) is total number of frames transmitted if the original frame must be transmitted i times
Go-Back-N ARQ Performance • For W>=2a+1, K is approximately equal to (2a+1) • For W<2a+1, K=W
High Level Data Link Control • HDLC is most important data link protocol • ISO 33009, ISO 4335 • Widely used and basis for many other important data link protocols